SECAI Core · phase 11 of 15 · noun
Data poisoning
When misinformation or malicious data is used in the training dataset, causing bias and changes in behavior
The Explain card
- Plain English
- Data poisoning plants misinformation or malicious samples in a training dataset so the resulting model learns bias, wrong answers, or hidden behaviors.
- Example
- An attacker edits a few hundred pages on a public wiki that a company scrapes for training. The model later confidently repeats the attacker's false claims about a competitor's product.
- Why it matters
- Training data is an attack surface that is huge, hard to audit, and often scraped from the open internet. Defenders track data provenance, filter sources, and test models for unexpected behaviors.
- Hook
- Poison the well and everyone downstream drinks it.
Word knowledge
How the word is built, where it came from, and what it sits beside in memory.
In a sentence
During validation, the security team traced a spike in false-negative intrusion alerts to data poisoning in the network-traffic training corpus.
Why these words
- data Latin datum, a thing given identifies the material affected
- poisoning Middle English poisonen, from Old French poisonner supplies the head noun and the corruption metaphor
Where it came from
- Origin
- English technical compound of data, from Latin datum, and poisoning, ultimately from Latin potio, drink or potion
- Entered the language
- 2000s
- What changed
- Poisoning broadened from administering a harmful substance to figurative corruption, then became associated with digital information and machine-learning security.
How it is spelled
- Pattern
- data preserves the Latin plural ending -a from datum
- Pattern
- poisoning adds -ing directly to poison without doubling the final consonant
Spelled like
- media
- criteria
- seasoning
- questioning
Broken into chunks
-
data
- database
- metadata
-
poison
- poisonous
- antipoison
-
-ing
- training
- learning
What it sits beside
Same subject
- adversarial machine learning
- training-data integrity
- backdoor attack
- model validation
Same shape
- cache poisoning
- DNS poisoning
- model poisoning
training-time attack vocabulary
- data poisoning
- backdoor attack
poisoning compounds in computing
- data poisoning
- cache poisoning
- DNS poisoning
Where it sits in the deck
Phase 11: Threat Landscape: Attack Vectors and Adversarial Techniques
With defences named, learn what they must defend against — the full catalogue of attack techniques targeting AI systems, their inputs, outputs, training pipelines, and supply chains.
